The NF-κB signal transduction pathway is the prototype of modular composites of functionally interdependent sets of proteins that coordinate to translate environmental stimuli into a cellular response [41]. In 2004, Bouwmeester et al. characterized the protein interaction network of Tumor necrosis factor (TNF)-α/NF-κB pathway components using an integrated approach comprising tandem affinity purification, liquid-chromatography tandem mass spectrometry, network analysis, and directed functional perturbation studies using RNA interference [41]. Such a mapping study identified FKBP51 for the first time as a major multifunctional kinase cofactor in NF-κB signaling among 221 molecular associations and 80 unknown interactors. This latter was co-purified with IκB kinase (IKK)α, IKKε, transforming growth factor-β-activated kinase 1 (TAK1), and mitogen-activated protein kinase kinase-1 (MEKK1). The interaction with IKKα was confirmed by co-immunoprecipitation. Following RNA interference of FKBP51, the authors demonstrated an impaired NF-κB activation, measured in a luciferase reporter assay, underlining an essential role for the immunophilin in the overall signaling process of NF-κB activation.

In 2014, Erlejman et al. [42] reported that FKBP51 impaired the nuclear translocation rate of NF-κB and its transcriptional activity. Because FKBP51 and FKBP52 are responsible in a mutually exclusive fashion for the retro-transport mechanism of steroid hormone receptors, the authors hypothesized that these immunophilins could similarly regulate the translocation of NF-κB proteins. They found an association of either FKBP51 or FKBP52 with NF-κB/RelA, with FKBP52 promoting the nuclear translocation while FKBP51 the cytoplasmic retention. Moreover, following Chromatin immunoprecipitation (ChIp) assays, they proposed that both immunophilins are recruited to the promoter sequences of NF-κB regulated genes, with FKBP51 inhibiting NF-κB transcriptional activity, while FKBP52 favors such activity.

In 2015, Romano et al. [43], by co-immunoprecipitation assays, confirmed the physical interaction between FKBP51 and IKKα and also demonstrated an interaction with IKKγ, IKKβ, and TNF-receptor associated factor 2 (TRAF2). FKBP51-knockdown inhibited the binding of IKKγ to the IKK catalytic subunits, i.e., IKKα and IKKβ, and impaired overall IKK catalytic activity. Either FKBP51 TPR and PPIase domains were required for their interaction with TRAF2 and IKKγ, whereas only the TPR domain was involved in interactions with IKKα and β. TRAF2 catalyzes K63-linked polyubiquitination of receptor-interacting protein 1 (RIP1) in response to TNFα receptor triggering; this non-canonical polyubiquitin chain recruits both TAK1 complex (consisting of TAK1 and TAK1-binding proteins TAB2, and TAB3) and the IKK complex, by binding directly to the ubiquitin-binding domains present on TAB2 and IKKγ. By interacting with TRAF2, FKBP51 reinforces IKKγ recruitment to the K63 ubiquitin chain that keeps closer RIP1, TAK1, and IKK [44]. FK506 and SAFits, respectively, unselective and specific inhibitors of the FKBP51 isomerase activity, impaired the IKK-regulatory role of FKBP51. These results support the conclusion that FKBP51 promotes NF-κB activation by serving as an IKK scaffold and an isomerase. Fig. 1 summarizes the different results from literature on FKBP51 in the NF-κB signaling.

A role for FKBP51 as a scaffold protein in AKT signaling was proposed in 2009 by Pei et al. when they identified the immunophilin as part of the protein complex AKT/PH domain leucine-rich repeat protein phosphatase (PHLPP), which enhanced the de-phosphorylation of AKT in a pancreatic cancer context [45]. The authors co-immunoprecipitated PHLPP1 and AKT with FKBP51, showing that the immunophilin, AKT, and PHLPP exist as a complex in cells. By employing FKBP51 mutants, they demonstrated that the N-terminus of the immunophilin (FK1 and FK2 domains) is bound to Akt, while the C-terminus, i.e., TPR domain, to PHLPP. They observed that, in the absence of FKBP51, Akt was hyperphosphorylated at Serine 473 (S473) due to the inefficient binding of PHLPP to Akt. Moreover, they found that AKT S473 phosphorylation was conserved upon the overexpression of the FKBP51 mutant FD67/68DV that lacks peptidylprolyl isomerase activity, suggesting that FKBP51 regulates AKT phosphorylation in an isomerase-independent manner and, thus, acting as a scaffolding protein to promote the AKT-PHLPP interaction [45]. Fabian et al. performed a detailed analysis of the domains involved in the interaction between FKBP51 and AKT, showing that AKT could directly bind to FKBP51 via the FK1 domain or indirectly via the TPR domain through heat-shock protein 90 (HSP90) [46]. Notably, other parts of the FK1 domain that do not exert PPIase activity are involved in the binding of FKBP51 to AKT [46]. Mutations in the FK1 site abolished the isomerase activity without affecting binding to AKT [46]. In 2017, the research group that first identified the role of FKBP51 as a scaffold protein for PHLPP [45] reported that this scaffold does not necessarily assist the de-phosphorylation of AKT [47]. They identified the sirtuin 7 (SIRT7) as an interactor of FKBP51 and responsible for its deacetylation at residues K28 and K155. They found that in different breast or prostate cancer cell lines, depletion of SIRT7 significantly increased the phosphorylation of AKT at S473 by regulating the acetylation status of the immunophilin [47]. Moreover, conversely, the FKBP51 acetylation was increased by P300/CBP and produced an increased AKT phosphorylation at S473 [47].

Gassen et al. reported a role for FKBP51 in autophagy regulation that involved AKT1 and PHLPP. They found a physical interaction between FKBP51 and Beclin-1 that increased the stability of Beclin-1 [48]. They showed that FKBP51 recruits E3 ubiquitin ligase S-phase kinase-associated protein 2 (SKP2) to Beclin-1. SKP2 executes K48-linked ubiquitination at K402 of Beclin-1, resulting in proteasomal degradation. The authors propose that the recruitment of SKP2 to FKBP51 produced defective K48-linked ubiquitination of Beclin-1 due to PHLPP-mediated AKT deactivation, which hampered protein degradation.

A very recent paper by our group demonstrated that the TPR domain of FKBP51 mediates AKT ubiquitination at K63, which is an essential step for AKT activation [49]. The spliced FKBP51, lacking such a domain, could not link K63-Ub residues to AKT. PHLPP stabilized the level of E3-ubiquitin ligase TRAF6 and supported K63-ubiquitination of AKT. This finding introduces an unknown oncogenic role for the phosphatase PHLPP supported by the interactome profile of FKBP51 carried out on melanoma cells overexpressing PHLPP [49]. An over-representation of components involved in cell division, transcription, and translation, along with a decrease of those related to apoptosis, highlighted a relevant role for PHLPP in improving oncogenic hallmarks. The spliced FKBP51 isoform can explain the discrepant results from the literature. Accelerated splicing towards the FKBP51s could shift the equilibrium between the two isoforms generating a reduced ability to ubiquitinate AKT due to the loss of the TPR domain. Also, a low tumor level of PHLPP, as it may occur in pancreatic cancer [49], can lead to a scarce activation of AKT on FKBP51-overexpression. Fig. 2 summarizes the different results from literature on FKBP51 in the AKT signaling.

Recent findings position FKBP51 as a central regulatory switch between AMPK and mTOR, pivotal proteins in autophagy regulation [48,50]. The balance between the adenosine 5′-monophosphate (AMP)–activated protein kinase (AMPK) and the mechanistic target of rapamycin (mTOR) regulates autophagy. WD-repeat protein interacting with phosphoinositides (WIPI) proteins are essential scaffold proteins that link autophagy’s key regulatory elements with metabolite-sensing enzymes [51]. Following Bakula et al., who demonstrated that WIPI3 and WIPI4 are essential scaffolders of the liver kinase B1 (LKB1)/AMPK/Tuberous Sclerosis (TSC)1/2 signaling network [51], Häusl et al. [50] found that FKBP51 recruits LKB1 to the WIPI4-AMPK regulatory platform to induce AMPK phosphorylation, which stimulates autophagy initiation by direct phosphorylation of UNC51-like kinase 1 (ULK1) complex, one of the most upstream components acting in the autophagy machinery. Furthermore, performing co-immunoprecipitation studies in neuronal cells, the authors showed an association of FKBP51 with TSC2, which depended on the presence of WIPI3, opening the possibility of an inhibitory role of FKBP51 in mTOR signaling. Fig. 3 illustrates the proposed mechanisms for FKBP51 roles in the AMPK/mTOR/autophagy axis.

Using a differential display technique, Giraudier et al. found increased expression of FKBP51 in megakaryocytes of patients suffering from idiopathic myelofibrosis [52]. The same group demonstrated that FKBP51 sustained the production and release of TGF-β1 in the bone marrow microenvironment of a murine model of myelofibrosis mediated by NF-κB activation [53]. These results highlighted for the first time a pathogenetic role for FKBP51 in such a myeloproliferative disorder characterized by fibrosis development [53]. The FKBP51 stimulating effect on TGF-β1 production was confirmed by Romano et al. in melanoma [54]. They also demonstrated a role for FKBP51 in regulating the TGF-β signaling [55]. By co-immunoprecipitation assays, Romano et al. found that FKBP51 interacts with the general transcriptional co-activator p300 and the TGF-β transcription factors Smad2/3 [55] and promotes some transcriptional activities of the TGF-β, precisely the gene expression of vimentin (VIM) and secreted protein acidic and cysteine-rich (SPARC), associated with accelerated tumor growth, invasion, and poor prognosis of melanoma [55]. ChIp assays showed FKBP51/p300 complexes bound to the promoter of the melanoma cancer stem-cell marker ATP-binding cassette transporter G2 (ABCG2) [54]. This finding, together with the observation of an increased transcript level of ABCG2 and an increased number of ABCG2+ cells upon FKBP51 overexpression in melanoma cells, further supported a role for FKBP51 as a co-regulator in gene expression [54]. Interestingly, ABCG2+ melanoma stem cells showed the highest expression levels of epithelial to mesenchymal transition (EMT) genes, including Twist basic helix-loop-helix transcription factor 1 (TWIST), Snail family of zinc-finger transcription factors (SNAIL), Snail Family Zinc Finger 2 (SLUG), CDH-2/N-cadherin, VIM, SPARC [54] highlighting EMT and cancer stemness are intertwined aspects in melanoma biology and progression.

By a cell-by-cell immunohistochemical analysis in colorectal cancer (CRC) and liver metastases resected after chemotherapy with oxaliplatin, Rotoli et al. showed a molecular interaction between the scaffold proteins angiomotin-like 2 (AmotL2), FKBP51, and IQ motif containing GTPase-activating protein 1 (IQGAP1) proteins within the EMT context of cancer tissues [56]. The colocalization of CD34 (a telocyte marker) with AmotL2, FKBP51, and IQGAP1 in tumor vessels was indicative of a role for these three scaffold proteins in tumor angiogenesis and vascular invasion [56]. They also found a tumor nest enveloped by several telocyte-like FKBP51+ cells. Notably, telocytes are interstitial stromal cells [57] and key players in regenerating and repairing organs. They establish a strict cell-cell interaction within the stem cell niches [58] by specific intercellular junctions. Functional interactions between AmotL2, FKBP51, and IQGAP1 proteins in several physiological and pathological situations were confirmed by String analysis showing the involvement of proto-oncogene SRC, HSP90, and yes-associated protein (YAP1) in AmotL2/FKBP51/IQGAP1 multicomplex [5]. Fig. 4 illustrates FKBP51 in the TGF-β signaling.

Several studies support the interaction of FKBP51 with the histone acetyltransferase HAT-p300 [47,54,55]. As mentioned in the AKT section, Yu et al. [47] observed a strict interaction among FKBP51 with p300 and CBP (CREB-binding protein) and found that FKBP51 acetylation was increased by CBP and, to a lesser extent, p300 [47]. In vitro assays showed that CBP and p300 acetylated FKBP51, while SIRT7 reverted this post-translational modification [47]. Notably, p300/CBP inhibition is currently explored as anticancer therapy [59,60] as it can affect the transcription of proteins involved in distinct oncogenic networks across different cancer types [59]. The interaction of FKBP51 with p300/CBP suggests an involvement of the immunophilin in chromatin modification [54]. Tufano et al. observed a general increase in HDACs and a reduced level of acetylated-p300 in FKBP51-KO melanoma cells [61]. These KO-cells also showed increased expression of the TNF-related apoptosis-inducing ligand (TRAIL) receptor DR5 and increased sensitivity to TRAIL-induced apoptosis. Death receptor 5 (DR5) expression is inhibited at the transcriptional level by the repressor activity of acetyl-Yin Yang 1 (YY1). YY1 is acetylated by p300 and deacetylated by HDACs [62]. A ChIP assay on KO-cells showed a reduced acetyl-YY1 on the DR5 promoter, associated with increased DR5 transcript levels. Reconstituting FKBP51 levels contrasted the effects of KO on DR5-, acetyl-YY1-, and acetyl-p300-levels [61]. Fig. 5 shows a graphical representation of the proposed relationship between FKBP51 and p300 and protein acetylation.

ChIp studies found FKBP51 bound to the CCND1 gene in an open chromatin status accompanied by increased CCND1 transcript levels [63]. This finding suggested a role for FKBP51 in participating in transcriptional complexes regulating the synthesis of cyclin D. By double immunofluorescence experiments on formalin-fixed and paraffin-embedded CRC samples, Rotoli et al. found FKBP51 and Proliferating cell nuclear antigen (PCNA) protein exhibiting nuclear colocalization in the early S phase in malignant cells forming tumor nests [56]. Cyclin D1 belongs to the core cell cycle machinery. Once induced, it binds and activates the cyclin-dependent kinase (CDK)4 and CDK6. Cyclin D-CDK4/6 holoenzymes phosphorylate proteins, such as the retinoblastoma protein (pRB), which governs cell cycle progression. In breast cancer, Jirawatnotai et al. identified FKBP51 as one of the most abundant CDK4 interacting proteins with a crucial role in CDK4 stability [64]. Newly translated CDK4 is first incorporated into a protein complex containing the chaperone HSP90 and the co-chaperone adaptor CDC37 in the cytoplasm, where CDK4 is stabilized in an inactive form [65,66]. The authors found that CDK4 interacts with FKBP51 or CDC37 but does not form a ternary CDK4-FKBP5-CDC37 complex, suggesting that FKBP51 stabilizes CDK4 level via a CDC37-independent mechanism. The authors found that depletion of FKBP51 decreased CDK4 protein levels and impaired CDK4 kinase activity, suggesting that FKBP51 promotes oncogenesis by stabilizing CDK4. In a mouse study, Ruiz-Estevez et al. [67] confirmed CDK4/FKBP51 interaction, but they found that FKBP51 sequesters CDK4 within the HSP90 storage complex and prevents the formation of the cyclin D1-CDK4 complex. FKBP51, through its isomerase activity, inhibited phosphorylation, hence activating CDK4. By these mechanisms, FKBP51 positively regulated myoblasts differentiation and myogenesis [67]. Fig. 6 illustrates the proposed mechanisms for FKBP51 role in cancer cell proliferation and chromatin remodeling.

Lagadari et al. demonstrated that FKBP51 interacts with human telomerase reverse transcriptase (hTERT) [68]. hTERT is the catalytic subunit of the enzyme telomerase, together with the telomerase RNA component (TERC), which comprises the essential unit of the telomerase complex [68]. As known, telomerase or terminal transferase is a ribonucleoprotein that adds a species-dependent telomere repeat sequence to the 3′ end of telomeres, thus contributing to protecting the end of the chromosome from DNA damage. hTERT is an HSP90 client-protein highly expressed in cancer cells, where it is required to compensate for the loss of telomeric DNA after each successive cell division. Lagadari et al. [68] found that FKBP51 is primarily localized in mitochondria, whereas hTERT is nuclear. Following oxidative stress, FKBP51 becomes nuclear and colocalizes with hTERT. FKBP51 promotes hTERT catalytic activity. The authors identified an essential role for FKBP51 PPIase activity for upregulating hTERT activity and for the TPR domain for efficient interaction of hTERT with HSP90 [68]. Fig. 7 illustrates FKBP51 role in telomerase elongation.

A role of FKBP51 in post-transcriptional regulation has been identified by Martinez et al., who found an association of FKPB5 to Argonaute 2 (Ago 2) in mouse embryonic stem cells (ESCs) [69]. Ago proteins load both siRNAs and miRNAs as short duplexes of ~22 bp; after this event, one strand, the guide, is stably retained while the other strand is degraded. Pharmacological inhibition of the Fkbp5-Ago2 interaction by the immunosuppressant FK506 or by Fkbp5 knockdown blocked miRNA-dependent stabilization of Ago2 expression resulting in decreased Ago2 protein expression, both in mouse and human cells. Ectopic FKBP5 expression increased Ago2 protein levels in a miRNA-dependent fashion. Given that FK506 binds to active PPIase domain, and FK506 treatment reduced Ago2 levels, along with the presence of two prolines in the RNA-bound Ago2 structure. A crucial role for Fkbp5 in Ago2 regulation was confirmed by mutagenesis analysis, indicating that both the PPIase and TPR domains facilitated Ago1 loading; however, the enzymatic function may be dispensable [70]. Accordingly, another FKBP, FKBP6, endowed with an inactive PPIase domain, plays a role in the biogenesis of Piwi-interacting RNAs [71]. Fig. 8 illustrates FKBP51 interaction with Ago2 and RNA processing.

Among the different protein chaperones that have been reported to either directly or indirectly influence the folding and expression of the programmed death ligand-1 (PDL-1) [72], FKBP51s plays a pivotal role in PDL-1 glycosylation and presentation on the cell surface [73]. PDL-1, also known as CD279 or B7-H1, is a 33-kDa type I transmembrane glycoprotein that contains IgV- and IgC-like domains in its extracellular region [74]. PDL-1 is physiologically expressed in macrophages, some activated T and B cells, dendritic cells (DCs), and some epithelial cells [75,76]. It can be expressed by tumor cells constitutively or adaptively in response to immune cell infiltration [77,78]. Although the molecular mechanism is still poorly characterized, it seems that FKBP51s acts at the level of the endoplasmic reticulum (ER) compartment where PDL-1 precursor protein is synthesized, thereby suggesting a possible role of the chaperone in the biogenesis of PDL-1 [73]. FKBP51s maintains the same structural organization as the full-length FKBP51 except for the lack of the protein-protein interaction domain TPR, which becomes replaced for the alternative splicing by a short new C-terminal amino acidic stretch whose function is still unknown [31]. However, the N-terminal portion of the chaperone consisting of the two PPIase domains (FK1 and FK2) remains intact and, as such, the peptidylprolyl cis/trans isomerase activity [73]. Accordingly, pharmacological inhibition of the PPIase activity by SAFit2 treatment affects PDL-1 glycosylation, plasma membrane expression, and contrasts PDL-1-induced cell death of activated lymphocytes cocultured with glioblastoma cells [73,79]. In an orthotopic mouse model of glioblastoma, daily treatment with SAFit2 significantly reduced tumor PDL-1 expression and growth [79]. Interestingly, PDL-1 belongs to the type I transmembrane protein family. Once the protein is correctly co-translationally translocated and inserted into the ER membrane, its topological organization provides that its proline-rich extracellular portion faces the ER lumen. In contrast, its short C-terminal tail, which does not contain any proline residues, faces the cytosolic side of the ER membrane (UniProt ID: Q9NZQ7). Due to the lack of a needful signal sequence to access the ER lumen, FKBP51s has a cytosolic localization which should not allow either interception or isomerization of proline residues of the PDL-1 polypeptide. D’Arrigo et al. reported that FKBP51s, and not the full-length FKBP51, was enriched in the ER fraction of glioblastoma cell lines suggesting a possible association with the ER membranes [73]. Much evidence has been shown that the ER entry of newly synthesized protein precursors can be assisted by different cytosolic proteins acting either at the level of the Sec61 translocon channel or the associated ribosomes during the co-translational translocation process [80–83]. In E. coli, a PPIase-containing chaperone, which plays a crucial role in assisting nascent chains destined for membrane translocation, is the trigger factor (TF) [84–87].

Interestingly, the PPIase domains of FKBP51s show somewhat structural similarity to the PPIase domain of TF. Indeed, as shown in Fig. 9, the 3D structure of the PPIase domain belonging to the TF displays a concave shape, which consists mainly of beta sheets with an alpha helix branch inside. This structural organization closely resembles the structure of the two PPIases domains of the FKBP51 protein, suggesting a possible similar function [84–87].

Notably, TF is a specialized chaperone supporting very early the protein folding program [87]. This ribosome-bound protein binds to short nascent chains as they emerge from the ribosome. Around the first 40 amino acids of the nascent chain are protected from the cytosol into the ribosomal exit tunnel [88,89]. Crosslink experiments have shown that TF binds to ribosomal proteins L23 and L29 in the exit tunnel [87]. In particular, the interaction with L23 has been reported as critical [87]. Even if TF possesses PPIase activity, the binding of TF to peptides is independent of the presence of proline residues [90]; therefore, the PPIase activity could not be required for the TF chaperoning of nascent chains.

Given the non-luminal ER localization of FKBP51s and the dependence of PDL-1 synthesis on the presence of a functional PPIase domain of FKBP51s, it could be conceivable to speculate that FKBP51s could assist PDL-1’s nascent chain as it emerges from the ribosome, thereby regulating its biogenesis during the co-translational translocation process driving the insertion of PDL-1 protein into the ER membrane. Fig. 10 illustrates the proposed mechanism for FKBP51s role in PDL-1 biogenesis.

A recent finding raises the hypothesis that FKBP51s participates, as the canonical FKBP51 in the events that govern cyclin-D oscillation being part of the transcriptional complex bound to CCND1 [63]. Following a study on mouse embryonic fibroblasts (MEFs) demonstrating a role for cyclin-D in the control of PDL-1 expression [91], a study conducted on glioblastoma cell lines showed that the level of PDL-1 changed during the cell cycle [63]. More precisely, PDL-1 mRNA appeared to increase concomitantly to CCND1 on G1/S transition, to decrease during exponential cell growth progressively. In the temporal window of PDL-1 and CCND1 peak, when PDL-1 protein expression level increased on the plasma membrane, FKBP51s localized in ER. On the decrease of cyclin-D and cell proliferation, FKBP51s went nuclear. We interrogated the CCND1 promoter and intronic sequences for their H3K27me3 and H3K4me3 pattern in A375 melanoma FKBP51-KO cells upon exogenous FKBP51 or FKBP51s expression. Our data revealed that H3K4me3 modifications, associated with high transcription activity, occurred in Flag-FKBP51 immunoprecipitated chromatin to the detriment of H3K27me3, thus producing a ratio of α-H3K4me3/α-H3K27me3 >1. In FKBP51s overexpressing cells, ChIp confirmed the binding to CCND1, but the α-H3K4me3/α-H3K27me3 arrangement was consistent with a closed chromatin conformation. In line with this finding, CCND1 mRNA levels were significantly reduced in FKBP51s overexpressing cells, compared to empty vector (EV) levels.